Advances in conventional total body irradiation (TBI) used for bone marrow transplant regimens have been stalled for five decades due to the inherent conflict between the efficacy of high dose irradiation (i.e., reduced relapse) and radiation induced toxicity. More specifically, studies show that higher doses of radiation reduce relapse, but increase toxicity to organs at risk (OAR) including the lungs, heart, eyes, liver, and kidneys. We propose to study the feasibility of a novel technique called adaptive total body and marrow irradiation (adaptive TBMI). This new approach has three clear advantages: 1) Incorporation of image guided tomotherapy that allows focused radiation to be delivered to the target (bone marrow and other disease sites), thereby differentially delivering doses of radiation to various organs; 2) Monitoring of radiation dose delivered to the patients and adjustment of subsequent treatments (as needed) to achieve the prescribed dose, also known as the adaptive process; and, 3) Allowance for higher radiation doses (dose escalation) without increasing toxicity by using an enhanced therapeutic ratio of dose to disease sites versus dose to OARs and soft tissues. These advantages will make it possible to conduct clinical trials to determine a safe and efficacious maximum tolerated dose (MTD) of TBMI in the setting bone marrow transplant. We will conduct a feasibility trial, using adaptive TBMI techniques to: (i) provide an understanding of body motion and the accuracy of dose delivery; (ii) individualize treatment through the adaptive processes; and, (iii) improving radiobiological precision of dose escalation. The dose escalation available through the enhanced therapeutic ratio of adaptive TBMI is expected to increase efficacy (i.e. leukemia kill) without increased toxicity to healthy organs. The central hypothesis of this work is that the dose escalation of adaptive TBMI is safe and efficacious, and provides a treatment option to patients with high risk hematological malignancies. We will test this hypothesis through two aims: 1) To Determine the maximum radiation dose of TBMI by performing a phase I dose escalation study and to estimate the efficacy of this approach in a phase II study, and 2) To optimize TBMI delivery by measuring the accuracy of 3D whole body localization within the scanner, measuring the accuracy of the TBMI dose delivery, and establishing an adaptive TBMI therapy process. Subjects (0-45 years of age) with advanced, chemotherapy refractory leukemia (those who fail to achieve complete remission) will be eligible. These patients have very poor survival, with most dying from their disease within weeks to months. If successful, TBMI may offer significant benefits over TBI through better leukemia control and thus, is expected to have a significant impact on patients with advanced leukemia and other hematologic diseases. Adaptive TBMI has the potential for better disease control, reduced disease recurrence and increased patient survival, consistent with the well-established NIH scientific mission.